All common RGB-D Saliency Datasets we collected are shared in ready-to-use manner.
We directly provide a ready-to-use download package, including all test sets, and two popular training sets in the RGB-D Saliency Detection.
- All test datasets, fetch code is b2p2.
- Training dataset 1, including NJUD and NLPR, fetch code is 76gu.
- Training dataset 2, including NJUD and NLPR, DUT, fetch code is 201p.
Datasets | NJUD | DUTLF-Depth | NLPR | SIP | STEREO | LFSD | RGBD135 | SSD | ReDWeb-S | COME15K |
---|---|---|---|---|---|---|---|---|---|---|
Size | 1985 | 1200 | 1000 | 929 | 797/1000 | 100 | 135 | 80 | 3179 | 15625 |
Publication | ICIP | ICCV | ECCV | TNNLS | CVPR | CVPR | ICIMCS | ICCVW | ReDWeb-S | ICCV |
Download link | here | here | here | here | here/1000_ori | here | here | here | here | here |
- Tips: All data is uniformly resized to 256×256 for training purposes.
- Please cite our paper if you use our dataset (DUTLF-Depth) in your research
@inproceedings{piao2019depth,
title={Depth-induced multi-scale recurrent attention network for saliency detection},
author={Piao, Yongri and Ji, Wei and Li, Jingjing and Zhang, Miao and Lu, Huchuan},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages={7254--7263},
year={2019}
}
Metrics | * | * | * | * |
---|---|---|---|---|
E-measure | S-measure | weighted F & F-measure | MAE | PR-curves |
Integration into a easy way. | * | * | * | The link is here |
- This Toolbox contains near all evaluation metrics for salient object detection including E-measure, S-measure, weighted F & F-measure, MAE scores and PR curves or bar metrics.
If you have any questions, please contact us ( [email protected] ).